Background. Human immunodeficiency virus (HIV)—associated nephropathy (HIVAN) is an important cause of end-stage renal disease among African American patients. This study was performed to study the epidemiology of HIVAN in a predominantly black African population and the impact of highly active antiretroviral therapy and other factors on the development of end-stage renal disease.
Methods. We retrospectively identified all patients with HIVAN, defined by biopsy or strict clinical criteria, in 8 clinics in the United Kingdom. Baseline renal function, HIV parameters, renal pathological index of chronic damage, and responses to highly active antiretroviral therapy were analyzed, and factors associated with adverse renal outcome were identified.
Results. From 1998 through 2004, we studied 16,834 patients, 61 of whom had HIVAN. HIVAN prevalence in black patients was 0.93%, and HIVAN incidence in those without renal disease at baseline was 0.61 per 1000 person-years. After a median of 4.2 years, 34 patients (56%) had developed end-stage renal disease. There were no significant differences in renal function and HIV parameters at baseline, time to initiation of highly active antiretroviral therapy, and rates of HIV RNA suppression between the 20 patients who developed end-stage renal disease >3 months after receiving the HIVAN diagnosis and the 23 patients who maintained stable renal function. However, the index of chronic damage score was significantly higher in those who developed end-stage renal disease (P<.001), and an index of chronic damage score >75 was associated with shorter renal survival (P<.001).
Conclusions. Whereas overall patient survival suggested an important benefit of highly active antiretroviral therapy, no additional renal benefit of early initiation of highly active antiretroviral therapy or viral suppression could be demonstrated in this large cohort of patients with established HIVAN. Severity of chronic kidney damage, as quantified by biopsy, was the strongest predictor of renal outcome.
HIV-associated nephropathy (HIVAN) is an important cause of chronic kidney disease (CKD) that has been described primarily in African American HIV-infected patients [1–3]. The reported prevalence of HIVAN in this population has a range of 3.5%, as determined by results of screens for proteinuria, to 12%, as determined by postmortem studies [4, 5]. Most studies of HIVAN originate in the United States, where HIVAN is associated with advanced immune suppression, a high proteinuria level, and rapid progression to end-stage renal disease (ESRD) [3, 6]. Before the availability of antiretroviral therapy (ART), the prognosis for patients with HIVAN was poor and was generally measured in weeks to months [7–9].
Several lines of evidence support a beneficial role of ART in improving the outcome and preserving the kidney function of patients with HIVAN. HIVAN incidence has declined during the HAART era , and uptake of ART has been associated with stabilization or improvement of renal function, reduction in the amount of proteinuria, and/or improvement of histological abnormalities, as described in case reports or small cohort studies [10–20]. In addition, 2 recent cohort studies of patients with HIVAN noted a beneficial effect of HAART and viral suppression on renal outcome [3, 6]. Nonetheless, the majority of patients with HIVAN in these studies progressed to ESRD.
In the United Kingdom, an increasing proportion of HIV-infected patients are black Africans . The prevalence of HIVAN among black African HIV-infected patients is unknown, and the natural history of HIVAN in this patient population remains to be defined in the presence of HAART. We undertook an epidemiological study of HIVAN among patients attending 8 large UK HIV clinics during the HAART era. We included patients with biopsy-confirmed HIVAN, as well as patients fulfilling strict clinical criteria ; calculated prevalence and incidence rates of HIVAN; and evaluated the effect of HAART, viral suppression, and the severity of biopsy-defined chronic kidney damage on renal outcome.
ase ascertainment and data collection All patients who received a diagnosis of HIVAN at 8 participating centers in the United Kingdom ([figure 1[figure 1) from January 1998 through December 2004 were retrospectively identified. Because not all patients with HIV and CKD underwent renal biopsy, we also included patients who fulfilled all of the following criteria: (1) estimated glomerular filtration rate (eGFR), <60 mL/min for >3 months; (2) proteinuria level, >1.5 g/24-h period; (3) absence of diabetes mellitus, hypertension, collagen vascular disease, liver cirrhosis, pregnancy, or organ transplant; and (4) ultrasound evidence of echogenic, unobstructed kidneys. The sensitivity and specificity of our case definition among 61 consecutive black patients with biopsy-confirmed kidney disease (45 with HIVAN; 16 with other etiologies) were 87% and 94%, respectively. Case ascertainment was performed through identification of all patients with biopsy-confirmed HIVAN, and systematic review of (1) all HIV-infected patients included in local renal databases and (2) all HIV-infected patients with proteinuria (at King's College, St. George's, North Middlesex, Royal Free, and Brighton hospitals only). Patients who had developed HIVAN before 1998 were excluded. During the study period, none of the participating HIV clinics performed urinalysis routinely. Approval for the study was obtained from the National Health Service Multi-centre Research Ethics Committee.
Demographic, clinical, laboratory parameters—including hypertension (use of antihypertensive medication or presence of severe hypertension [systolic blood pressure ⩾160 mm Hg and/or diastolic blood pressure ⩾100 mm Hg]) and diabetes mellitus [glucose intolerance requiring dietary or pharmacological interventions]) at baseline—all available CD4+ T cell counts, plasma HIV RNA levels, serum creatinine levels, blood pressures, HIV-treatment regimens, and information about the subsequent clinical course—including use of cortocosteroids, angiotensin-converting enzyme inhibitors, and/or angiotensin-receptor II blockers, start and stop dates of dialysis, and status at the last clinic visit—were abstracted from electronic databases or the medical records and were last updated in June 2007.
Assessment of renal function and definition of renal outcome Renal function was assessed by eGFR, which was calculated using the 4-variable modification of diet in renal disease equation . ESRD was defined as stage 5 CKD (eGFR <15 mL/min or initiation of permanent renal replacement therapy). Patients were stratified into 3 categories according to renal outcome: (1) ESRD <3 months after HIVAN diagnosis, (2) ESRD >3 months after HIVAN diagnosis, and (3) stable native renal function. Demographic, clinical, and laboratory parameters were described for all patients with HIVAN, for the subsets of patients with biopsy-confirmed or clinically defined HIVAN, and for those with specified renal outcomes. Because all patients who initiated ART received a nonnucleoside reverse-transcriptase inhibitor or a (boosted) protease inhibitor in combination with 2 nucleoside reverse-transcriptase inhibitors, the term “HAART” refers to “dual class, triple combination therapy.” Viral suppression was defined as ⩾3 consecutive HIV RNA levels <50 copies/mL during a period of ⩾6 months, and viral rebound was defined as ⩾2 subsequent consecutive HIV RNA levels >400 copies/mL after having attained viral suppression.
Review of renal pathology Renal biopsy specimens, where available, were reviewed by 2 pathologists blind to renal outcome. Specimens were identified that showed collapsing glomerulopathy (collapse of capillary loops with prominence of visceral epithelial cells). Tubules showed different amounts of atrophy, particularly seen as cystic dilatation with flattened epithelium . Patients whose biopsy specimens showed other glomerular abnormalities, including noncollapsing focal segmental glomerulosclerosis (5 patients), were excluded. The amount of chronic kidney damage was measured on sample sections stained by periodic acid-methenamine silver, if sample sections were available. Areas of cortex showing global sclerosis of glomeruli, tubular atrophy, and interstitial fibrosis were outlined on computer images, and their area was measured and expressed as a percentage of cortical area to give an index of chronic damage (ICD) score, which can have a range of 0%–100%, as described elsewhere .
Data analysis Prevalent HIVAN cases were defined as renal failure (eGFR, <60 mL/min) or proteinuria level >1.5 g/24-h period at the time of HIV diagnosis, and incident cases were defined as the development of HIVAN among patients who presented with normal renal function (eGFR of ⩾60 mL/min and no documented proteinuria) at the time of HIV diagnosis. Survival was calculated using Stata software, version 8 (Stata Corp.), from the time of HIV diagnosis (prevalent cases) or the date the eGFR was first <60 mL/min, first documented proteinuria (>1.5 g/24-h period), or renal biopsy, whichever occurred earliest (incident cases). Renal survival was calculated as time from HIVAN diagnosis or from the date of renal biopsy to (1) initiation of permanent renal replacement therapy or (2) reaching stage 5 CKD. If patients did not require renal replacement therapy and the last available eGFR was ⩾15 mL/min, renal survival was censored at the time of the last visit or death. Survival curves were compared by log-rank test. Categorical variables were compared by Fisher's test of exact probability, and continuous variables were compared by Wilcoxon's rank-sum or Student's t test.
Prevalence and incidence of HIVAN Among 16,834 patients who attended the 8 participating HIV centers, we identified 58 patients with HIVAN. No HIVAN cases were identified among 10,403 nonblack (predominantly white) patients. In 48 patients (83%), HIVAN was present at the time of HIV diagnosis. HIVAN prevalence was 0.93% (95% CI, 0.67%–1.19%) among 5147 (black) patients who received a new diagnosis of HIV infection during the study period, whereas the HIVAN incidence was 0.61 (95% CI, 0.33%–1.13%) per 1000 person-years among 6383 black patients who attended participating HIV centers during the study period and who did not have HIVAN at baseline. The 10 patients who developed HIVAN, although known to be HIV infected, had defaulted from HIV care for a median of 3.5 years (range, 2–10 years) (6 patients), were experiencing ART failure (3 patients), or were not receiving ART (1 patient) at the time of the HIVAN diagnosis. An additional 3 patients with HIVAN from participating renal centers who received HIV care in clinics where HIV care was indistinguishable from that of other centers but where information technology systems prevented us from performing a systematic search (Manchester Hospital and Chelsea and Westminster Hospital) were also included in the study.
haracteristics of patients with HIVAN The diagnosis of HIVAN was confirmed on renal biopsy for 45 patients (74%), and 16 patients (26%) who did not have renal biopsies performed fulfilled the clinical case definition ([figure 1[figure 1). Data about proteinuria status were available for 54 (89%) of these 61 patients, and proteinuria level was >1.5 g/24-h period in 52 (96%) of the 54 patients. There were no statistically significant differences in demographic, clinical, or laboratory parameters between patients with biopsy-confirmed HIVAN and with clinically defined HIVAN (table 1). In contrast to previous studies of African American patients with HIVAN [3, 6], the prevalence of viral hepatitis, diabetes mellitus, and hypertension (coincident with HIVAN diagnosis) in our largely black African patient population was low (overall, <10%). Overall, 46 (77%) of 60 HIV-1—infected patients attained an undetectable HIV RNA level, and a single patient with biopsy-confirmed HIVAN was infected with HIV-2.
linical outcome Overall patient survival was 89% at 5 years (figure 2a) and did not differ between patients with biopsy-confirmed and clinically defined HIVAN (P=.51). By June 2007, there were 34 patients (56%) who had developed ESRD (including 4 patients with stage 5 CKD who had not yet started dialysis). Median renal survival was ∼3.5 years (figure 2b) and did not differ between patients with biopsy-confirmed and clinically defined HIVAN (P=.87). Among 45 patients with histologically confirmed HIVAN, the median time to ESRD, calculated from the date of renal biopsy, was ∼3 years (figure 2c).
To assess the relationship among HAART, viral suppression, CD4+ T cell recovery, and renal outcome, patients were stratified according to renal outcome (see Methods). To prevent possible misclassification, we excluded 3 patients with progressive renal failure who had not yet developed ESRD and 3 patients with initially stable renal function in whom HAART interruption and rebound viremia were associated with rapid progression to ESRD (figure 3).
Patients who developed ESRD <3 months after receiving the HIVAN diagnosis had ESRD (median eGFR, 5 mL/min) and a high proteinuria level (median, 8.2 g/24-h period) at initial presentation. Interestingly, patients who progressed to ESRD >3 months after receiving the HIVAN diagnosis and those who maintained stable renal function had similar renal function at the time of HIVAN diagnosis (median eGFR, 28 and 29 mL/min, respectively), and similar proportions of patients commenced HAART within the first 3 months or attained viral suppression or CD4+ T cell counts >200 cells/µL within the first year after HIVAN diagnosis (table 2). When this analysis was restricted to patients with biopsy-confirmed HIVAN, similar results were obtained ([table 3[table 3).
The relationship between viral suppression and renal function over time was also studied in each patient individually. In 9 patients, renal function improved considerably after resolution of presenting infections and correction of volume depletion before HAART initiation. Although viral suppression was noted in 74% of patients with stable renal function, 65% of patients who progressed to ESRD >3 months after HIVAN diagnosis had HIV RNA levels <50 copies/mL for a median duration of 691 days (interquartile range, 501–1177 days) at the time of initiating dialysis. The 3 patients who experienced progressive decrease in renal function but who had not yet developed ESRD (figure 3) all had undetectable HIV RNA levels for >1000 days at their last clinic visit, whereas renal function remained stable in 6 patients, despite persistent viremia for a median of 839 days.
Recurrent viremia was observed in 11 patients, 4 of whom had ESRD at the time of loss of viral suppression. In 2 patients, renal function was unaffected by viral recurrence; in 2 patients, renal function was transiently reduced by 40%–50%, whereas in 3 patients, in whom renal function had remained stable (eGFR for 2 patients, ∼45 mL/min; eGFR for 1 patient, ∼20 mL/min) for 7–26 months while HIV replication was suppressed, progression (within 3 months) to ESRD was noted when HAART was discontinued and HIV RNA levels increased.
Pathological analysis Renal biopsies of samples from 29 HIV-1—infected patients with HIVAN were available for review. Seventeen biopsy samples had been obtained within 1 month after HIVAN diagnosis, and 12 had been obtained after a median of 183 days (interquartile range, 66–342 days). All had characteristic features of HIVAN. The median ICD score for 16 patients who developed ESRD was 84 (interquartile range, 78–88), compared with 31 (interquartile range, 18–68) for 12 patients who maintained stable renal function (P<.001) (1 patient with progressive renal failure had an ICD score of 60). The severity of chronic damage did not differ for the 9 and 7 patients who progressed to ESRD <3 and >3 months, respectively, after the biopsy date (median ICD score, 80 vs. 85; P=.31). Even after exclusion of 6 patients who presented with ESRD <3 months after HIVAN diagnosis, patients who progressed to ESRD had significantly higher ICD scores than did those with stable renal function (median ICD score, 83 vs. 31; P=.004). Renal survival was significantly longer for patients with ICD scores <75 than for patients with ICD scores >75 (P<.001) (figure 2d), despite similar rates of viral suppression (77% and 75%, respectively). ICD scores >75 had a sensitivity of 87.5%, specificity of 84.6%, positive predictive value of 87.5%, and negative predictive value of 84.6% for development of ESRD.
Our study demonstrates that, although HAART has dramatically improved survival of patients with HIVAN, renal survival remains poor, even for patients who have attained complete viral suppression. The majority of patients in our study had advanced renal failure at the time of HIVAN diagnosis. In these patients, the amount of chronic kidney damage, as measured on renal biopsy, was the factor most strongly associated with renal outcome. We did not demonstrate an association between early HAART initiation, rapid viral suppression, or CD4+ T cell recovery and renal outcome. Our results suggest that, once severe HIVAN is established (ICD score, >75), HAART and viral suppression are unlikely to prevent progression to ESRD.
Case reports and small retrospective studies have suggested a beneficial effect of ART on renal outcome in patients with HIVAN [10–15, 18], although follow-up was generally short. One study noted stable renal function in 5 patients who received HAART and attained viral suppression and progressive renal failure in 6 patients not receiving HAART , whereas another study noted better renal survival in 26 patients with HIVAN who received HAART . A further study of 42 patients with HIVAN also noted slower progression to ESRD in patients receiving ART and an association between viral nonsuppression and development of ESRD . However, few patients attained viral suppression, and ∼80% progressed to ESRD within 3 years of renal biopsy [3, 6]. In our study, 77% of patients attained viral suppression, and median renal survival was somewhat better than in the African American patients [3, 6]. More widespread use of HAART and higher rates of viral suppression may have contributed to better renal outcome, as may have differences in patient characteristics, including the absence of injection drug use and the low prevalence rates of diabetes, hypertension, and viral hepatitis in our largely black African patient population.
Our study highlights the fact that the association between viral suppression and renal function is complex and confounded by varying degrees of chronic kidney damage. Renal function may improve substantially before HAART initiation, may remain stable in patients with persistent viral replication, or may decline progressively in patients with prolonged viral suppression. HAART discontinuation and rebound viremia may adversely affect kidney function , as shown in 3 of our patients who rapidly progressed to ESRD after HAART discontinuation. The findings for these patients and a fourth patient reported in the literature  suggest that HAART and/or viral suppression may be critically important in preserving renal function in some patients with HIVAN. The extremely low incidence of HIVAN (0.25 per 1000 patient-years) among black patients “in care” in the United Kingdom provides further support that there is an important role for HAART in the prevention of or delay in the clinical manifestations of HIVAN .
Our study has several limitations. None of the participating centers' HIV units routinely performed urinalysis. It is likely, therefore, that mild HIVAN cases remained undetected. This may have resulted in an underestimation of HIVAN prevalence in patients with newly diagnosed HIV infection and an overestimation of the proportion of patients with HIVAN who had severe renal failure and/or had progressed to ESRD. Given the fact that HAART was generally offered to patients who presented with advanced HIV infection and that HAART may favorably affect the natural history of subclinical HIVAN , the effect of a lack of regular screening for proteinuria on the detected incidence of clinical HIVAN in our patients was probably limited.
A further limitation is that not all patients underwent renal biopsy. Proteinuria and renal failure in black African patients may be caused by a number of renal pathologies that are clinically indistinguishable from HIVAN, including HIV-immune complex kidney disease . Given the fact that all HIV and renal parameters at baseline and the natural history were indistinguishable between patients with histologically confirmed and clinically defined HIVAN and that <10% of black patients who fulfilled the HIVAN case definition and underwent renal biopsy had diagnoses other than HIVAN, we are confident that most patients with clinically defined HIVAN did indeed have HIVAN. Moreover, we obtained similar results when the analysis was restricted to patients with biopsy-confirmed HIVAN.
Follow-up for patients who did not progress to ESRD was shorter than for patients who developed ESRD >3 months after HIVAN diagnosis. Some patients who were classified as having stable renal function may thus have progressed to ESRD with longer follow-up. However, given the sample size of our study and the prolonged duration of follow-up, the effect of potential misclassifications on survival estimates would have been relatively small. In addition, we limited the likelihood of misclassification by excluding from the comparative analysis 3 patients who experienced a gradual decrease in renal function without reaching ESRD and by classifying as having ESRD 4 patients with an eGFR <15 mL/min who had not initiated permanent renal replacement therapy at their most recent visit.
In summary, HIVAN affects ∼1% of black patients who receive HIV care in the United Kingdom. Optimal management of HIV infection, including initiation of HAART when CD4+ T cell counts reach 200–350 cells/µL , is associated with a very low incidence of HIVAN. All black patients with HIV infection should be screened for HIVAN by urinalysis at baseline and annually thereafter if they are not receiving or not responding to HAART, in keeping with Infectious Diseases Society of America/HIV Medicine Association guidelines . Early detection of HIVAN allows prompt initiation of HAART to reduce morbidity and mortality associated with immune suppression and to reduce the risk of progressive kidney damage in those with early HIVAN lesions. HAART interruptions should be avoided in patients with HIVAN and, given the excellent survival rates of patients with HIVAN who respond to HAART, renal transplantation should be considered for those with viral suppression, CD4+ T cell counts >200 cells/µL, and ESRD due to HIVAN .
We thank Katrina Casley-Ready, Anthony Warrens, Stephen Duffell, David Waluube, Phillip Hay, Catherine Macks, Kevin Jones, Anele Waters, Anthi Balitsari, Heather Noble, Nicky Perry, Anthony Pullin, Achim Schwenk, Brian Gazzard, and Colette Smith, for their assistance with information technology, research and development submission, access to hospital notes, supplying data, and/or organizing site participation; Candice Roufosse, for assistance with the pathology review; pathologists at participating institutions, for providing sections for review; and Fowzia Ibrahim, for statistical advice.
Financial support Guy's and St. Thomas' Charity (R041041 [to F.A.P.]).
Potential conflicts of interest All authors: no conflicts.